From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756435AbcJWLFv (ORCPT ); Sun, 23 Oct 2016 07:05:51 -0400 Received: from mx1.redhat.com ([209.132.183.28]:34230 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751993AbcJWLFs (ORCPT ); Sun, 23 Oct 2016 07:05:48 -0400 Date: Sun, 23 Oct 2016 13:05:42 +0200 From: Jiri Olsa To: Ingo Molnar Cc: Arnaldo Carvalho de Melo , linux-kernel@vger.kernel.org, Linux Weekly News , Andi Kleen , David Ahern , Don Zickus , Jiri Olsa , Joe Mario , Namhyung Kim , Peter Zijlstra , Arnaldo Carvalho de Melo Subject: Re: [GIT PULL 00/52] New Tool: perf c2c Message-ID: <20161023110542.GA15618@krava> References: <1476975876-2522-1-git-send-email-acme@kernel.org> <20161022082806.GA4526@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20161022082806.GA4526@gmail.com> User-Agent: Mutt/1.7.1 (2016-10-04) X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.38]); Sun, 23 Oct 2016 11:05:47 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Oct 22, 2016 at 10:28:06AM +0200, Ingo Molnar wrote: SNIP > > tools/perf/ui/browsers/hists.h | 1 + > > tools/perf/util/hist.c | 1 + > > tools/perf/util/hist.h | 1 + > > tools/perf/util/mem-events.c | 128 ++ > > tools/perf/util/mem-events.h | 37 + > > tools/perf/util/sort.c | 2 +- > > tools/perf/util/sort.h | 1 + > > 13 files changed, 3210 insertions(+), 2 deletions(-) > > create mode 100644 tools/perf/Documentation/perf-c2c.txt > > create mode 100644 tools/perf/builtin-c2c.c > > Pulled the perf-c2c-for-mingo-20161021 tag, thanks a lot Arnaldo! > > I can see some teething problems. For example if I run it on an older kernel (v4.4 > distro kernel), I get this: > > triton:~/tip> perf c2c record perf bench sched pipe > # Running 'sched/pipe' benchmark: > # Executed 1000000 pipe operations between two processes > > Total time: 12.001 [sec] > > 12.001919 usecs/op > 83320 ops/sec > [ perf record: Woken up 18 times to write data ] > [ perf record: Captured and wrote 5.356 MB perf.data (69804 samples) ] > > but there's no 'perf c2c report' TUI output at all: > > Shared Data Cache Line Table (0 entries, sorted on remote HITMs) > Total Rmt ----- LLC Load Hitm ----- ---- Store Reference ---- --- Load Dram ---- LLC Total ----- Core Load Hit ----- -- LLC Load Hit - > Index Cacheline records Hitm Total Lcl Rmt Total L1Hit L1Miss Lcl Rmt Ld Miss Loads FB L1 L2 Llc Rm > > and just an empty screen. > > If I do 'perf report' I get two events: > > Available samples > 24K cpu/mem-loads,ldlat=30/P > 45K cpu/mem-stores/P > > and both have some real data. > > What am I missing? hm, most likely you did not generate any remote HITMs.. if it was on the laptop? you might get some results for local HITMs with -d option: $ perf c2c report -d lcl the default display/filter is for Remote HITMs.. perhaps we could be smart enough and switch to local when we found no data for remote jirka